Tuning Equilibrium Compositions in Colloidal Cd1–xMnxSe

Dec 7, 2015 - The physical properties of semiconductor nanocrystals can be tuned dramatically via composition control. Here, we report a detailed inve...
0 downloads 7 Views 1MB Size
Tuning Equilibrium Compositions in Colloidal Cd1−xMnxSe Nanocrystals Using Diffusion Doping and Cation Exchange Charles J. Barrows, Pradip Chakraborty, Lindsey M. Kornowske, and Daniel R. Gamelin* Department of Chemistry, University of Washington, Seattle, Washington 98195-1700, United States S Supporting Information *

ABSTRACT: The physical properties of semiconductor nanocrystals can be tuned dramatically via composition control. Here, we report a detailed investigation of the synthesis of high-quality colloidal Cd1−xMnxSe nanocrystals by diffusion doping of preformed CdSe nanocrystals. Until recently, Cd1−xMnxSe nanocrystals proved elusive because of kinetic incompatibilities between Mn2+ and Cd2+ chemistries. Diffusion doping allows Cd1−xMnxSe nanocrystals to be prepared under thermodynamic rather than kinetic control, allowing access to broader composition ranges. We now investigate this chemistry as a model system for understanding the characteristics of nanocrystal diffusion doping more deeply. From the present work, a Se2−-limited reaction regime is identified, in which Mn2+ diffusion into CdSe nanocrystals is gated by added Se2−, and equilibrium compositions are proportional to the amount of added Se2−. At large added Se2− concentrations, a solubility-limited regime is also identified, in which x = xmax = ∼0.31, independent of the amount of added Se2−. We further demonstrate that Mn2+ in-diffusion can be reversed by cation exchange with Cd2+ under exactly the same reaction conditions, purifying Cd1−xMnxSe nanocrystals back to CdSe nanocrystals with fine tunability. These chemistries offer exceptional composition control in Cd1−xMnxSe NCs, providing opportunities for fundamental studies of impurity diffusion in nanocrystals and for development of compositionally tuned nanocrystals with diverse applications ranging from solar energy conversion to spin-based photonics. KEYWORDS: nanocrystals, quantum dots, diffusion doping, cation exchange, giant Zeeman effect, magnetic circular dichroism

N

semiconductors, making this ion unsuitable for doping into the internal volumes of II−VI nanocrystals via cation-exchange chemistries,29 and the high temperatures that would be required to accelerate its diffusion have widely been associated with Mn2+ out-diffusion from nanocrystals (“self-purification”).30,31 We found that by tuning Mn2+ chemical potentials in solution to be greater than those inside the NCs, Mn2+ in-diffusion into various II−VI semiconductor nanocrystals could be made thermodynamically favorable, but diffusive composition scrambling still requires long reaction times at elevated temperatures, for which deliberate suppression of Ostwald ripening is also essential if the desired nanostructuring of the preformed host crystallites is to be retained.25 To this end, we demonstrated that Ostwald ripening can be effectively eliminated by raising the chemical potentials of all cations in solution using a large excess of solvated Mn2+.25 Among existing synthetic methods for making Cd1−xMnxSe nanocrystals, diffusion doping has yielded simultaneously the largest Mn2+ content and the greatest size uniformity, making NCs prepared by this method especially attractive for spectroscopic investigations and subsequent applications. Cd1−xMnxSe

anocrystal (NC) composition control represents a central challenge in nanoscience. Although there has been tremendous progress in recent years,1−10 many desirable compositions still remain unattainable in colloidal NCs, motivating the development of new chemistries. One important area of composition control pertains to semiconductor nanocrystal doping11−16 because the addition of impurities can dramatically change the electronic, magnetic, or optical properties of the NCs, imparting desirable physical properties that in turn may stimulate fundamental discoveries or applied technologies. Understanding the fundamental chemical boundaries of nanocrystal composition control for new and challenging dopant/host combinations could yield unprecedented doped semiconductor nanomaterials for applications from spectral conversion in lighting and luminescent solar concentrators (LSCs)17,18 to optical nanothermometry,19 bioimaging,20 plasmonics,21−23 or spin-based electronic/photonic information processing.24 Recently, we reported a “diffusion doping” approach for introducing relatively incompatible and slowly diffusing cations into semiconductor nanocrystals.25 This approach involves thermodynamically driven cation + anion addition followed by hightemperature diffusive cation randomization. Cd1−xMnxSe was selected as our model system because of the well-known and longstanding challenges of synthesizing high-quality Cd1−xMnxSe nanocrystals.24,26−29 Mn2+ diffusion is notoriously slow in II−VI © XXXX American Chemical Society

Received: September 24, 2015 Accepted: December 7, 2015

A

DOI: 10.1021/acsnano.5b07389 ACS Nano XXXX, XXX, XXX−XXX

Article

www.acsnano.org

Article

ACS Nano nanocrystals prepared by diffusion doping have now been used to demonstrate rich upper-excited-state magnetic circular dichroism (MCD),32 extraordinarily large excitonic Zeeman splittings,33 excitonic magnetic polarons,34 and enhanced positive magnetoresistance suggestive of electron magnetic polarons.35 Although the products of this chemistry are already proving useful, many details of the chemistry itself remain unexplored. For example, previous work has not systematically explored relationships between synthesis conditions and product compositions, and tunable equilibrium compositions have not been demonstrated. Further studies that better characterize this promising synthetic method should help accelerate its adaptation to other even more challenging target compositions. At the same time, further experimentation will provide new fundamental insights into ion diffusion in semiconductor nanostructures in general, and will also inform our understanding of the stabilities of doped nanocrystals under high-temperature processing or device operation conditions. Here, we report the results of synthetic experiments aimed at improving our understanding of CdSe nanocrystal diffusion doping with Mn2+. Key new observations include the demonstration of complete compositional equilibration, widely tunable equilibrium compositions via added Se2−, existence of a thermodynamic Mn2+ solubility limit, and reversal of the diffusion doping via cation exchange. These new findings highlight several unique features of this chemistry that allow its use for extraordinarily fine composition control in doped semiconductor nanocrystals, advancing the frontier of doped nanocrystal synthesis and elevating the quality of doped nanocrystals accessible for future physical, spectroscopic, or technological applications.

Figure 1. Room-temperature (A) electronic absorption and (B) 1.5 T MCD spectra of undoped d = 3.7 nm CdSe nanocrystals (0.1 mmol, black), the same nanocrystals after 20 min in the presence of Mn2+ at 300 °C (0.1 mmol, red), and 24 h following the addition of Se2− (0.04 mmol, blue). These concentrations correspond to a ratio of 0.4:1:1 Se2−:Mn2+:Cd2+ ions in CdSe (referred to here as Se2−:Mn2+:CdSe). Spectra for other aliquots are omitted for clarity. (C) Time evolution of gExc during diffusion doping to form Cd1−xMnxSe NCs, as determined by analysis of the MCD spectra of these and other aliquots.

solvated Mn2+ (vide inf ra, Figure 2). In this particular experiment, injection of Se2− into this reaction mixture after 20 min at 300 °C causes a blue shift of the first exciton by ∼80 meV and inversion and intensification of its associated MCD intensity in the 24 h aliquot. Both of these spectroscopic changes are signatures of Mn2+ incorporation into the internal volumes of the CdSe nanocrystals. Addition of Mn2+ into CdSe nanocrystals introduces strong sp−d magnetic exchange interactions between the unpaired Mn2+(3d) spins and the exciton. The sp−d contribution to ΔEZeeman is parametrized by the last term of eq 2a, where xeff represents the concentration of paramagnetic Mn2+ (reduced from the true x because of Mn2+−Mn2+ antiferromagnetic superexchange interactions38,39) γ describes the overlap between the ensemble of Mn2+ dopants and the excitonic wave function (γ = 1 for uniform Mn2+ distribution), ⟨SZ⟩ is the Mn2+ spin expectation value at the experimental temperature and magnetic field, and N0(α−β) describes the mean-field sp−d exchange energy. In the Curie limit, this sp−d term converges to the linear term gsp−dμBB, such that gExc = gint + gsp−d (eq 2b). Because N0(α−β) = −1.5 eV for Mn2+-doped CdSe, gsp−d < 0 and Mn2+ doping causes the sign inversion observed in ΔEZeeman (Figure 1).13,24,28,37,40 In this way, room-temperature (Curie limit) MCD spectroscopy can be used as a sensitive analytical probe of Mn2+ incorporation into the CdSe nanocrystals.

RESULTS Reactions in our initial report25 on diffusion doping were all performed using a large excess of added Se2−. A primary goal of the present study is to clarify the role of the added Se2− in this chemistry. Figure 1 summarizes spectroscopic data collected during a reaction specifically designed to illustrate two key features of nanocrystal diffusion doping: doping gated by Se2−, and equilibration. Figure 1A,B shows room-temperature electronic absorption and MCD spectra of three aliquots taken during this reaction: (i) the initial seed CdSe nanocrystals, (ii) the same nanocrystals after 20 min heating at 300 °C in the presence of a large excess of Mn2+ but without added Se2−, and (iii) after 24 h heating following injection of additional Se2−. The absorption and MCD spectra of the seed CdSe nanocrystals show multiple resolved excitonic transitions, consistent with a narrow size distribution. The first excitonic absorption maximum occurs at ∼2.15 eV, implying an average nanocrystal diameter of d = 3.7 nm.36 The MCD spectrum of the same sample shows a derivative-shaped A-term feature coincident with the first excitonic absorption maximum, arising from the Zeeman splitting of this excitonic excited state. The negative leadingedge intensity of this derivative signal is indicative of a positive excitonic Zeeman splitting energy (ΔEZeeman > 0) and hence a positive intrinsic excitonic g value (gint > 0, eq 1, where μB is the Bohr magneton and B is the magnetic field).37 Quantitative analysis13,33 of these spectra yields a value of gExc = gint ≈ +1 for the seed CdSe nanocrystals. ΔEZeeman = gExcμBB

ΔEZeeman = g int μB B + xeff γ ⟨SZ⟩N0(α − β) = (g int + gsp−d )μB B (Curie limit)

(1)

(2a) (2b)

Figure 1C summarizes the MCD results from several aliquots taken from this reaction, plotting gExc versus reaction time in a semilogarithmic representation. Prior to Se2− addition, gExc

The absorption and MCD spectra of these nanocrystals remain unchanged after many hours at 300 °C, despite the large excess of B

DOI: 10.1021/acsnano.5b07389 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

yields a universal time constant of τeff ≈ 1.3 h that reproduces the entire data set reasonably well. Under these conditions, diffusion doping thus proceeds with roughly the same very slow kinetics at different Se2− loadings, reaching different equilibrium values of x in the resulting Cd1−xMnxSe nanocrystals. Under the conditions of our previous report,25 reactions run with excess Se2− showed evolving differences between spectroscopic and total (ICP-AES) Mn2+ contents, demonstrating evolving Mn2+ concentration gradients within the NCs prior to equilibration, thereby implicating slow Mn2+ lattice diffusion as the rate-determining process. Because the diffusion-doping kinetics appear independent of added Se2− (Figure 2A), we conclude that very slow Mn2+ diffusion through the NC lattice is also rate-determining in the Se2−-limited regime. Figure 2B summarizes equilibrium data from a broad series of diffusion-doping reactions spanning from very small amounts of added Se2− up to a large excess of added Se2− (25:1:1 Se2−:Mn2+:CdSe), with all other reaction conditions held constant. In addition to equilibrium |gsp−d| values, Figure 2B plots Mn2+ concentrations (x in Cd1−xMnxSe) determined by two independent methods: analytically (ICP-AES), and via analysis of the exciton blue shift according to the known dependence of Eg on x in bulk Cd1−xMnxSe,40 extrapolated to the nanoscale.25 The two measures of x agree well, and both also track the magnitude of |gsp−d| determined spectroscopically. The equilibrium values of x rise sharply with added Se2− up to ∼5 equiv but then plateau at x = 0.31 ± 0.03, at which point |gsp−d| also plateaus at ∼26. The reaction thus exhibits two qualitatively different regimes: At low Se2− concentration, diffusion doping is Se2−-limited, whereas at high Se2− concentration, x is limited by other factors (discussed below). The data in Figures 1 and 2 demonstrate that this diffusion doping involves equilibrium control over nanocrystal compositions, that is, compositions controlled thermodynamically rather than kinetically.25 As an equilibrium process, we hypothesized that this chemistry should be reversible. Specifically, under competition conditions, Cd2+ should out-compete Mn2+ for lattice Se2− in Cd1−xMnxSe nanocrystals because of the highly favorable Cd2+−Se2− thermodynamics relative to Mn2+−Se2− (bond energies: ∼310 vs ∼201 kJ/mol at 298 K41). To test this hypothesis, a diffusion-doping reaction was performed to incorporate Mn2+ into CdSe NCs under Se2−-limited conditions (0.5 Se2−:CdSe) in the presence of a large excess of Mn2+, as described above. After equilibration, Cd2+ was then added to the same reaction mixture (without introducing additional Se2−, T = 280 °C; see Methods). Figure 3A shows absorption spectra of aliquots taken from this reaction. During diffusion doping, the first exciton blue shifts by 51 meV (from 2.076 to 2.127 eV) at equilibrium, consistent with Mn2+ incorporation to reach x ∼ 0.13 (Figure 2B).25 Upon Cd2+ addition, the first exciton red shifts again by 115 meV to 2.012 eV at equilibrium, overshooting its original energy. The change in exciton energies from before diffusion doping to after cation exchange suggests that the nanocrystals have grown from d = ∼4.4 to ∼5.4 nm, again indicating that x ∼ 0.13 in the intervening Cd1−xMnxSe nanocrystals. These results illustrate that diffusion doping proceeds by addition of Mn2+−Se2− units and hence involves nanocrystal growth in accordance with the number of Mn2+ ions added, but subsequent nanocrystal purification proceeds by cation exchange with no net change in the total number of cations. TEM analysis of aliquots removed from a similar reaction confirms that the nanocrystals grow during diffusion doping but remain nearly the same size after cation exchange (see Supporting Information).

Figure 2. (A) Time evolution of |gsp−d|, as determined by roomtemperature MCD, from undoped d = 3.9 nm CdSe NCs with varying amounts of excess Se2− (relative to Cd2+ in CdSe) during diffusion doping at 300 °C. The lines show results from a global fit of all of the data to a single exponential function, yielding τeff = 1.3 h. All reactions were performed at a ratio of 1:1 added Mn2+ to Cd2+ in CdSe(1:1 Mn2+:CdSe). Electronic absorption and MCD spectra for representative aliquots are shown in the Supporting Information. (B) Se2− dependence of xMn, determined from analysis of absorption spectra (red triangles) and analytically determined by ICP-AES (green squares), after 24 h of diffusion doping at 300 °C with 0−25:1:1 Se2−:Mn2+:CdSe. The |gsp−d| values (blue circles) for these samples are plotted on the right axis and scaled to match the trend shown in xMn. The dashed line is a guide for the eye. Reactions performed multiple times have yielded values within ∼5% of the mean.

remains unchanged. After Se2− addition, gExc shifts smoothly from its initial value of gExc = gint ≈ +1 to a new equilibrium value of gExc = gint + gsp−d ≈ −10. The room-temperature sp−d exchange contribution is thus described by gsp−d ≈ −11 for this sample. The data in Figure 1 demonstrate that Mn2+ incorporation into these CdSe nanocrystals is gated by the addition of Se2−, consistent with our previous interpretation that diffusion doping involves stoichiometric cation + anion addition.25 Furthermore, whereas most reactions in our previous report were run for only ∼30 min, these data show that equilibrium is not actually reached for several hours after Se2− addition. Next, we wished to explore the dependence of this reaction on the amount of Se2− added. Figure 2A plots the evolution of gsp−d measured during several ∼24 h diffusion-doping reactions performed under identical conditions except using different amounts of added Se2−. As in Figure 1, if no Se2− is added, then gExc does not change even after several hours, indicating that gsp−d = 0 at all times and hence the CdSe NCs remain undoped. Adding Se2− induces Mn2+ doping and consequently an increase in |gsp−d|. For these reactions, addition of more Se2− leads to a greater equilibrium value of |gsp−d|. Each reaction equilibrates over ∼6−8 h with similar kinetics. The curves in Figure 2A were obtained from a global fit of all data points to a single exponential function. Use of an exponential fitting function is helpful for capturing the phenomenological effective reaction rates but is not meant to imply a particular reaction mechanism. This fitting C

DOI: 10.1021/acsnano.5b07389 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 3. (A) Room-temperature electronic absorption spectra of undoped d = 4.4 nm CdSe NCs (black) diffusion doped at 280 °C with 0.5:1:1 Se2−:Mn2+:CdSe (blue traces) and purified at 280 °C by cation exchange (red traces), triggered after 20 h by injection of cadmium oleate (1.92 Cd2+ added per Mn2+ dopant). (B) Room-temperature MCD spectra of undoped CdSe (black), diffusion-doped Cd1−xMnxSe after 20 h (blue), and the same NCs after 6 h of cation exchange (red). Spectra for additional aliquots are omitted for clarity. (C) EPR spectra of representative aliquots, with solid line spectra matching the times of the MCD spectra shown in panel B. (D) Normalized Gaussian widths (σ) of the first excitonic absorption (green triangles, from analysis of panel A data), |gsp−d| (blue circles, from analysis of panel B data), and quantitative integrated EPR intensities (red squares, from analysis of panel C data), over the course of diffusion doping and cation exchange. The black lines are from global single exponential fits to the spectroscopic signatures of diffusion doping and cation exchange. The syringe cartoon indicates when Cd2+ was added to the reaction mixture.

In addition to these energy shifts, the first exciton’s Gaussian line width increases from σ = 44 to 62 meV during diffusion doping, and remarkably, it decreases back to 44 meV again after cation exchange. The increase in σ with Mn2+ doping is thus not a result of Ostwald ripening, but instead, it reflects additional inhomogeneous broadening in the Cd1−xMnxSe NCs, presumably coming from statistical and spatial Mn2+ distributions within the NC ensemble, as well as Mn2+ magnetic fluctuations.42 Finally, Figure 3B plots MCD spectra of the initial CdSe NCs, the equilibrium diffusion-doped Cd1−xMnxSe NCs, and the same nanocrystals equilibrated again after Cd2+ addition. The first excitonic MCD feature inverts and grows in intensity during diffusion doping but diminishes again after Cd2+ addition, reverting to a signal consistent with undoped CdSe NCs. These data demonstrate that Cd2+ addition to the reaction mixture under diffusion-doping conditions indeed reverses the Mn2+ doping, purifying the nanocrystals despite the presence of a large excess of solvated Mn2+. The absorption and MCD spectroscopic probes of Mn2+ doping are both selective for Mn2+ ions within the NC cores, where Mn2+−exciton interactions are strong. Furthermore, analysis of these data yields the effective concentrations that would be required to generate the same spectroscopic observables in the limit of uniform distribution of Mn2+ throughout the entire NC volume. It is conceivable that these techniques might not faithfully report on the total Mn2+ content of the nanocrystals, for example, if Mn2+ were concentrated near the nanocrystal surfaces. For this reason, EPR spectroscopy was also used to follow the progress of the same diffusion doping and purification reaction described above. EPR spectroscopy provides a measure of the paramagnetic Mn2+ content, independent of Mn2+ position within the NCs. A limitation of this EPR experiment is that it can be performed only after thorough washing and resuspension of the NCs to remove the excess solvated Mn2+ because this procedure also likely removes

surface-bound Mn2+. Nevertheless, this experiment provides a valuable independent measure of Mn2+ content. Figure 3C plots EPR spectra of the same aliquots described above and shows intensification of a broad g = 2 feature consistent with Mn2+ in Cd1−xMnxSe nanocrystals during diffusion doping.25 The Mn2+ hyperfine structure is not resolved because of the high Mn2+ concentrations involved (xmax ∼ 0.13). The EPR intensity diminishes and disappears after addition of Cd2+, consistent with full expulsion of Mn2+ from the nanocrystals. Figure 3D summarizes the complete collection of data from this entire set of measurements, plotting the evolution of σ, |gsp−d|, and the integrated EPR intensity. For convenience of comparison, each metric has been normalized within its undoped (0) and maximum diffusion-doped (1) range. All three spectroscopic signatures follow the same trends during the diffusion-doping and purification steps. In the purification step of the experiment described in Figure 3, we added ∼1.9 equiv of Cd2+ per lattice Mn2+. To investigate the lattice purification chemistry more broadly, a series of reactions was performed under identical conditions but with different amounts of added Cd2+. Figure 4A plots |gsp−d| versus reaction time for several of these reactions. The black trace reproduces the data from Figure 3. The green trace shows that adding a substoichiometric amount of Cd2+ (0.24 Cd2+ per lattice Mn2+) yields similar equilibration kinetics but a smaller decrease in |gsp−d|, terminating at |gsp−d| ∼40% of its initial value; that is, part but not all of the Mn2+ is removed by a substoichiometric Cd2+ addition. The blue traces show data from a similar experiment involving two sequential injections of 0.6 Cd2+ equivalents each, separated by 4 h. After the first Cd2+ addition, |gsp−d| decreases and equilibrates at ∼35% of its initial value. After the second addition of Cd2+, |gsp−d| decreases again and equilibrates at 0. A control experiment in which oleic acid and ODE were added without Cd2+ caused no meaningful change in gsp−d. Use of Na+ instead of Cd2+ also does not change gExc. These control D

DOI: 10.1021/acsnano.5b07389 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

of the last Mn2+ ions from the NCs requires some excess of Cd2+. These trends may in part reflect statistical distributions of both Mn2+ and added Cd2+ over the ensemble of nanocrystals, or they could reflect a changing ratio of equilibrium constants for Mn2+−Se2− versus Cd2+−Se2− bond formation. Addition of even more Cd2+ (up to 12 equiv; see Supporting Information) does not cause any further change in the absorption or MCD spectra, indicating that further Cd2+ is not incorporated into the nanocrystals after the Mn2+ has been replaced.

DISCUSSION Scheme 1 summarizes the diffusion-doping and cation-exchange chemistries reported here. In the presence of a large excess of Scheme 1. Mn2+ Diffusion Doping of CdSe Nanocrystals and Mn2+ Expulsion from Cd1−xMnxSe Nanocrystals via Cation Exchange with Cd2+ Figure 4. (A) Time evolution of |gsp−d| during cation exchange of equilibrated, diffusion-doped Cd1−xMnxSe nanocrystals held at 280 °C. In gray pentagons, no Cd2+ was added. In green squares and black circles, the Cd2+ (0.24 and 1.92 added Cd2+:lattice Mn2+, respectively) was added all at once. In blue triangles, 1.20 added Cd2+:lattice Mn2+ was divided equally into two injections, and the NCs were allowed to equilibrate after each injection of 0.60 added Cd2+:lattice Mn2+. Solid lines are a global fit to a single exponential with τeff = 1.1 h. The dashed curve is a guide for the eye. (B) Dependence of the extent of purification by cation exchange on added Cd2+, monitored as |gsp−d| (normalized) of equilibrated samples after cation exchange from diffusion-doped NCs, plotted as a function of the number of equivalents of Cd2+ added relative to the amount of Mn2+ in the diffusion-doped NCs. The red “+” and purple “×” markers are from successive titrations of Cd2+ in two separate reaction mixtures, allowing equilibration before each subsequent titration. Each other color/symbol pair represents the equilibrium value obtained from panel A. The dashed curve is a guide for the eye. Electronic absorption and MCD spectra of representative reaction aliquots are provided in the Supporting Information.

solvated Mn2+, nanocrystal doping with Mn2+ is gated by the addition of Se2− to the reaction mixture. The quantity of added Se2− tunes the composition of the resulting Cd1−xMnxSe nanocrystals by dictating how much Mn2+ is transferred from solution to the nanocrystal at equilibrium (eq 3a, Se2−-limited regime). The nanocrystals’ composition in turn tunes their corresponding magneto-optical properties. Because diffusion doping involves addition of Mn2+ + Se2− to the nanocrystal lattice, the nanocrystal diameter increases during diffusion doping. After diffusion doping, the addition of Cd2+ to the reaction mixture under the same conditions triggers cation exchange, expelling Mn2+ ions from the lattice and replacing them with the added Cd2+ ions (eq 3b). The cation-exchange reaction preserves the total number of cations and anions within the nanocrystal but tunes the composition. Both reactions are equilibrium processes.

experiments demonstrate that it is indeed the Cd2+ addition that causes nanocrystal purification of Mn2+, consistent with the proposed cation-exchange process. Interestingly, the purification kinetics are similar for all added Cd2+ concentrations. The solid curves in Figure 4A represent the results of a global fit of all reaction kinetics to a single exponential function, which yields τeff = 1.1 h. Even at its slightly lower temperature (280 vs 300 °C), this purification chemistry is marginally faster than the diffusion doping (τeff = 1.3 h, Figure 2), but both are remarkably slow. The similarity of the diffusiondoping and cation-exchange reaction rates is consistent with both kinetics being determined by very slow Mn2+ diffusion through the CdSe NC lattice. Figure 4B summarizes equilibrium values of |gsp−d| measured after equilibration of several different NC purification reactions under the same conditions as in Figure 3, plotted versus equivalents of added Cd2+. The data are consistent with |gsp−d| returning to 0 at ∼1 equiv of Cd2+ per lattice Mn2+. The curvature of these data is interesting, however, and not fully understood. Initially, the slope of the curve exceeds 1 Mn2+ lost from the NC per added Cd2+, suggesting that Cd2+ addition may initiate diffusion of more than one Mn2+, perhaps by aiding the formation of a steady-state concentration of Frenkel defects. At the other end of the curve, however, the slope decreases below 1 Mn2+ lost from the NC per added Cd2+, indicating that removal

Diffusion doping: CdSe + y Mn 2 + + ySe 2 − ⇌ CdMnySe1 + y (= Cd1 − xMnxSe) (3a)

Cation exchange: CdMn ySe1 + y + yCd2 + ⇌ Cd1 + ySe1 + y (= CdSe) + y Mn 2 + (3b)

With small quantities of added Se , x in the final Cd1−xMnxSe NCs is Se2− limited. The Cd1−xMnxSe equilibrium composition thus changes in direct proportion to the amount of added Se2−. Above x ∼ 0.31, however, x becomes independent of added Se2−. In this regime, no further precipitation of Mn2+ occurs in any form, even when a 25-fold excess of Se2− is introduced. Curiously, the slope of x versus added Se2− in Figure 2B is significantly smaller than expected from the stoichiometry of eq 3a, suggesting that a large fraction of the added Se2− is somehow unreactive. We speculate that this Se2− might be absorbed in the form of soluble (Mn−Se)n clusters. This lost Se2− does not reappear upon addition of excess Cd2+, however, perhaps suggesting that it is unreactive for another reason such as incomplete in situ reduction of the metallic Se precursor. Despite extensive efforts, we have found no evidence of MnSe 2−

E

DOI: 10.1021/acsnano.5b07389 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Scheme 2. Cd1−xMnxSe Lattice Formation Energies and the Normalized Concentrations of Reaction Components Plotted versus Composition and Added Se2−

Scheme 3. Microscopic Contributions to Nanocrystal Diffusion Doping

below the analogous limit for crystals grown from the ionic melt (xmax ∼ 0.4 at 300 °C, extrapolated from the CdSe−MnSe phase diagram40), where there is no solvent and there are no ligands to stabilize Mn2+ outside of the growing CdSe-based crystal. This comparison illustrates the dependence of xmax on the chemical potentials of the ions in the liquid phase and simultaneously emphasizes the very high Mn2+ content achievable by this nanocrystal diffusion-doping method. Microscopically, the role of the added Se2− during diffusion doping is two-fold. The addition of Se2− to the nanocrystal surfaces is formally equivalent to introducing surface cation vacancies (Scheme 3), which lowers the chemical potential (μMn) of lattice Mn2+ relative to solvated Mn2+, allowing it to become thermodynamically favorable for Mn2+ from solution to bind to the surface Se2− and become part of the nanocrystal lattice. The cation vacancies formed upon Se2− addition also aid cation diffusion within the lattice, which is necessary for scrambling Mn2+ positions to reach the randomized equilibrium Cd1−xMnxSe compositions observed experimentally. The diffusivity of Mn2+ in bulk II−VI semiconductors is strongly enhanced when the lattice is nonstoichiometric.43 Nonstoichiometries accelerate cation diffusion by increasing the concentrations of Frenkel-type cation vacancies, which play a central role in the Mn2+ diffusion mechanism. In nanocrystals, the surface cation vacancies (excess surface Se2− anions) can be filled by subsurface cations, thereby generating lattice vacancies (VCd, Scheme 3) that accelerate cation diffusion throughout the internal nanocrystal volume, for example, via a cation vacancy− interstitial (Frank−Turnbull) mechanism.44−46 The correlation between magneto-optical intensities and analytical Mn 2+ concentrations indicates that Mn2+ enters the lattice randomly and not via a preferred sector, a discrete shell, or layer-by-layer substitution along specific crystallographic planes, as often observed in cation-exchange reactions.5−8,47 Under the same reaction conditions as used for diffusion doping, addition of Cd2+ removes Mn2+ from the Cd1−xMnxSe NCs. When Cd2+ is added, it is thermodynamically able to

precipitation under any of our reaction conditions (in any crystallographic or amorphous phase), even in control experiments performed in the absence of CdSe nanocrystals. This latter observation is interpreted as indicating that MnSe is too soluble to precipitate under these conditions, tentatively attributed to the presence of coordinating amines and oleic acid in our reaction mixtures. Given the high solubility of MnSe and the insolubility of CdSe under our reaction conditions, it follows that there must exist some threshold value of x above which Cd1−xMnxSe also cannot be precipitated. The experimental data indicate that this solubility limit occurs at xmax ∼ 0.31 under our reaction conditions. The equilibrium compositions thus transition from being Se2−-limited to being solubility-limited as the amount of added Se2− is increased. Scheme 2A summarizes these observations by plotting the free energy of formation (ΔGf) for the Cd1−xMnxSe nanocrystals versus the composition parameter x. Up to x ∼ 0.31, ΔGf < 0 under our reaction conditions, but for x > ∼0.31, ΔGf > 0, and consequently nanocrystals with x > xmax = 0.31 are not observed. Because of the insolubility of crystallites with x ≤ xmax, Cd1−xMnxSe nanocrystals heated in the presence of excess Se2− and excess Mn2+ reach but do not exceed xmax. It follows that the equilibrium concentration of solvated Se2− is small in the Se2−-limited regime but grows in proportion to added Se2− in the solubility-limited regime (Scheme 2B). Similarly, the equilibrium concentration of solvated Mn2+ decreases linearly with added Se2− in the Se2−-limited regime but becomes independent of added Se2− in the solubility-limited regime, mirroring the Mn2+ uptake by the nanocrystals. Under our experimental conditions, ∼55% of all Mn2+ remains in solution in the solubility-limited regime after maximum diffusion doping. The specific value of xmax obviously must depend on the specific reaction conditions, such as temperature, ligand identity and concentration, Mn2+ concentration, and possibly even nanocrystal size or faceting, but a thorough exploration of this vast parameter space was considered beyond the scope of the present investigation. Nevertheless, it is interesting to note that our xmax lies only a little F

DOI: 10.1021/acsnano.5b07389 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano outcompete Mn2+ for its coordinated Se2− anions, which has the effect of raising μMn in the lattice relative to μMn in solution and thereby driving Mn2+ into solution. Analogous to composition tuning during diffusion doping via added Se2−, nanocrystal compositions can be finely tuned via the amount of added Cd2+ during cation exchange. When added approximately stoichiometrically, Cd2+ displaces all Mn2+ from the nanocrystals. Unlike many other nanocrystal cation-exchange processes, however, the exchange of Cd2+ for Mn2+ is extremely slow, requiring several hours at 280 °C instead of seconds or less at room temperature.3,48 To prevent Ostwald ripening during this slow cation exchange, μCd must remain substantially lower in the lattice than in solution, a condition achieved through the presence of excess solvated Mn2+. This extremely slow cation exchange is interpreted as reflecting specifically a very slow diffusion of Mn2+ in CdSe, even in the nanocrystalline form. In stoichiometric Mn2+ δ-doped CdTe wafers, Mn2+ diffusion measured between 420 and 530 °C shows Arrhenius behavior described by a temperature-independent Mn2+ diffusion constant (D0) of 1.7 × 10−6 cm2/s and an activation energy of EA = 1.35 eV.43 Extrapolation of these data down to our reaction temperature of 280 °C suggests that Mn2+ should require ∼22 h to diffuse 2.7 nm, the average radius of our nanocrystals. By some measures of bulk Mn2+ diffusivity,49 this diffusion should even take an order of magnitude longer. Although slow, the data in Figure 4 thus suggest diffusivities in the nanocrystals that still exceed those in bulk by at least an order of magnitude and possibly more, consistent with the general notion of enhanced ion diffusion in nanocrystals.3,46 Although the precise microscopic origins of this enhancement are not yet clear, and many possible explanations exist invoking surface nonstoichiometries or interfacial strain, it is undoubtedly associated with surface proximity; even in bulk, accelerated cation diffusion near surfaces appears to be a general phenomenon.43 Notably, the Mn2+ in- and out-diffusion kinetics do not differ markedly from one another in our experiments (Figure 3D), suggesting a relatively narrow range of Mn2+ diffusivities even with the different surface chemistries of these two processes. With kinetic data of the types shown in Figures 2 and 4 now available for the Mn2+ diffusion-doping and cation-exchange reactions discussed here, modeling studies, perhaps in conjunction with atomistic reaction-coordinate calculations, should make possible an even deeper understanding of the microscopic diffusion pathways involved in these interesting chemistries. Further work in this direction is presently underway.

compositions of Cd1−xMnxSe nanocrystals can therefore be tuned continuously from 0 ≤ x ≤ ∼ 0.31 in two ways: (i) during diffusion doping by changing the amount of added Se2−, or (ii) during cation exchange by changing the amount of added Cd2+. Combining diffusion-doping and cation-exchange methodologies thus leads to unprecedented composition control in Cd1−xMnxSe NCs. All of these chemistries are performed under conditions in which Ostwald ripening has been suppressed, allowing this composition control in preformed high-quality colloidal CdSe nanocrystals prepared separately by independent state-of-the-art synthetic methods. The advances in compositional tuning reported here enable further studies of microscopic cation diffusion processes and of advanced dopant-dependent physical properties in semiconductor nanocrystals, which in turn will enable development of new high-quality colloidal doped nanocrystals for target applications ranging from solar energy conversion to spintronics.

METHODS Synthesis. Oleate-capped seed CdSe nanocrystals were prepared by methods adapted from refs 36 and 50−52, as detailed in ref 25. Typically, diffusion doping of CdSe nanocrystals with Mn2+ was performed as follows: CdSe nanocrystals (∼0.1 mmol in terms of CdSe units) were dried and added to 1 mL of octadecene (ODE), 1 mL of tributylphosphine (TBP), and the desired amount (0−2.5 mmol) of selenium powder, in a septum-capped 5 mL round-bottom flask in a nitrogen-atmosphere glovebox. Separately, 12 g of ODE, 0.5 g of oleic acid (OA), and 1 g of hexadecylamine (HDA) were added to a 100 mL three-neck round-bottom flask. Following heating of the latter solution for 60 min at 100 °C under vacuum, 0.025 g (0.1 mmol) of Mn(OAc)2·4H2O was added against a nitrogen overpressure. The flask was then placed under vacuum to remove acetic acid and water and then heated under nitrogen to 300 °C, at which point the CdSe/selenide solution was injected rapidly. This reaction mixture was held at 300 °C and allowed to equilibrate for between a few seconds and 1 day. In NC purification experiments, diffusion doping was performed at 280 °C for 20−24 h. A solution of cadmium oleate (0−0.25 mmol) dissolved in 2 mL of ODE was then dripped into the flask, and the system was allowed to re-equilibrate for up to 1 day at 280 °C. After equilibrium was reached, the solution was cooled to room temperature and washed by repeated suspensions in toluene and flocculation with ethanol. In these experiments, no freshly nucleated CdSe or MnSe particles were ever detected by absorption, luminescence, TEM, or XRD measurements. Physical Characterization. Relative atomic concentrations were determined by analysis of dried nanocrystals digested in ultrapure nitric acid (EMD Chemicals) using inductively coupled plasma atomic emission spectrometry (ICP-AES; PerkinElmer). CW electron paramagnetic resonance (EPR) experiments were performed on equally concentrated colloidal toluene suspensions of nanocrystals using an X-band Bruker EMX spectrometer. Room-temperature electronic absorption spectra were collected for colloidal toluene suspensions of nanocrystals in a 0.1 cm path length cuvette using a Cary 500 spectrophotometer. Room-temperature magnetic circular dichroism spectra were collected on the same cuvette placed in a 1.5 T electromagnet oriented in the Faraday configuration. MCD spectra were collected using an Aviv 40DS spectropolarimeter. The differential absorption collected in the MCD experiment is reported as ΔA = AL − AR, where AL and AR refer to the absorption of left and right circularly polarized photons in the sign convention of Piepho and Schatz.53,54 From these data, values of ΔEZeeman and gExc can be obtained.13,33,40

CONCLUSION Diffusion doping allows nanocrystals to achieve their thermodynamically favored compositions without reaching their thermodynamically favored shapes or sizes.25 Here, we have investigated in detail the model system of Cd1−xMnxSe nanocrystals. A Se2−limited reaction regime is identified, in which Mn2+ diffusion into CdSe nanocrystals is gated by the addition of excess Se2−. In this regime, the Mn2+ content (x) at equilibrium is proportional to the amount of added Se2−. A second, solubility-limited regime is also identified, which occurs for large values of added Se2−. In this regime, x at equilibrium is limited to xmax ∼ 0.31 by the thermodynamic instability of Cd1−xMnxSe nanocrystals that have x > xmax. Consequently, x at equilibrium is independent of the amount of added Se2− in this regime. After diffusion doping, Mn2+ in-diffusion can be reversed by cation exchange with Cd2+, effectively purifying Cd1−xMnxSe nanocrystals back to CdSe nanocrystals. Under these reaction conditions, the equilibrium

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.5b07389. Additional absorption and MCD spectra and TEM data (PDF) G

DOI: 10.1021/acsnano.5b07389 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

(18) Bradshaw, L. R.; Knowles, K. E.; McDowall, S.; Gamelin, D. R. Nanocrystals for Luminescent Solar Concentrators. Nano Lett. 2015, 15, 1315−1323. (19) McLaurin, E. J.; Bradshaw, L. R.; Gamelin, D. R. Dual-Emitting Nanoscale Temperature Sensors. Chem. Mater. 2013, 25, 1283−1292. (20) Wang, F.; Banerjee, D.; Liu, Y.; Chen, X.; Liu, X. Upconversion Nanoparticles in Biological Labeling, Imaging, and Therapy. Analyst 2010, 135, 1839−1854. (21) Lounis, S. D.; Runnerstrom, E. L.; Llordés, A.; Milliron, D. J. Defect Chemistry and Plasmon Physics of Colloidal Metal Oxide Nanocrystals. J. Phys. Chem. Lett. 2014, 5, 1564−1574. (22) Scotognella, F.; Della Valle, G.; Srimath Kandada, A.; ZavelaniRossi, M.; Longhi, S.; Lanzani, G.; Tassone, F. Plasmonics in HeavilyDoped Semiconductor Nanocrystals. Eur. Phys. J. B 2013, 86, 1−13. (23) Routzahn, A. L.; White, S. L.; Fong, L.-K.; Jain, P. K. Plasmonics with Doped Quantum Dots. Isr. J. Chem. 2012, 52, 983−991. (24) Beaulac, R.; Archer, P. I.; Ochsenbein, S. T.; Gamelin, D. R. Mn2+Doped CdSe Quantum Dots: New Inorganic Materials for SpinElectronics and Spin-Photonics. Adv. Funct. Mater. 2008, 18, 3873− 3891. (25) Vlaskin, V. A.; Barrows, C. J.; Erickson, C. S.; Gamelin, D. R. Nanocrystal Diffusion Doping. J. Am. Chem. Soc. 2013, 135, 14380− 14389. (26) Mikulec, F. V.; Kuno, M.; Bennati, M.; Hall, D. A.; Griffin, R. G.; Bawendi, M. G. Organometallic Synthesis and Spectroscopic Characterization of Manganese-Doped CdSe Nanocrystals. J. Am. Chem. Soc. 2000, 122, 2532−2540. (27) Magana, D.; Perera, S. C.; Harter, A. G.; Dalal, N. S.; Strouse, G. F. Switching-on Superparamagnetism in Mn/CdSe Quantum Dots. J. Am. Chem. Soc. 2006, 128, 2931−2939. (28) Archer, P. I.; Santangelo, S. A.; Gamelin, D. R. Inorganic Cluster Syntheses of TM2+-Doped Quantum Dots (CdSe, CdS, CdSe/CdS): Physical Property Dependence on Dopant Locale. J. Am. Chem. Soc. 2007, 129, 9808−9818. (29) Erwin, S. C.; Zu, L.; Haftel, M. I.; Efros, A. L.; Kennedy, T. A.; Norris, D. J. Doping Semiconductor Nanocrystals. Nature 2005, 436, 91−94. (30) Dalpian, G. M.; Chelikowsky, J. R. Self-Purification in Semiconductor Nanocrystals. Phys. Rev. Lett. 2006, 96, 226802. (31) Chen, D.; Viswanatha, R.; Ong, G. L.; Xie, R.; Balasubramaninan, M.; Peng, X. Temperature Dependence of “Elementary Processes” in Doping Semiconductor Nanocrystals. J. Am. Chem. Soc. 2009, 131, 9333−9339. (32) Fainblat, R.; Muckel, F.; Barrows, C. J.; Vlaskin, V. A.; Gamelin, D. R.; Bacher, G. Valence-Band Mixing Effects in the Upper-Excited-State Magneto-Optical Responses of Colloidal Mn2+-Doped CdSe Quantum Dots. ACS Nano 2014, 8, 12669−12675. (33) Barrows, C. J.; Vlaskin, V. A.; Gamelin, D. R. Absorption and Magnetic Circular Dichroism Analyses of Giant Zeeman Splittings in Diffusion-Doped Colloidal Cd1−xMnxSe Quantum Dots. J. Phys. Chem. Lett. 2015, 6, 3076−3081. (34) Nelson, H. D.; Bradshaw, L. R.; Barrows, C. J.; Vlaskin, V. A.; Gamelin, D. R. Picosecond Dynamics of Excitonic Magnetic Polarons in Colloidal Diffusion-Doped Cd1‑xMnxSe Quantum Dots. ACS Nano 2015, 9, 11177−11191. (35) Liu, H.; Guyot-Sionnest, P. Magnetoresistance of ManganeseDoped Colloidal Quantum Dot Films. J. Phys. Chem. C 2015, 119, 14797−14804. (36) Yu, W. W.; Qu, L.; Guo, W.; Peng, X. Experimental Determination of the Extinction Coefficient of CdTe, CdSe, and CdS Nanocrystals. Chem. Mater. 2003, 15, 2854−2860. (37) Archer, P. I.; Santangelo, S. A.; Gamelin, D. R. Direct Observation of sp-d Exchange Interactions in Colloidal Mn2+- and Co2+-Doped CdSe Quantum Dots. Nano Lett. 2007, 7, 1037−1043. (38) Furdyna, J. K.; Kossut, J. Diluted Magnetic Semiconductors; Academic Press: New York, 1988; Vol. 25 of Semiconductors and Semimetals. (39) Shapira, Y.; Foner, S.; Ridgley, D. H.; Dwight, K.; Wold, A. Technical Saturation and Magnetization Steps in Diluted Magnetic

AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This research was funded by the U.S. National Science Foundation (DMR-1408617 to D.R.G.). C.J.B. acknowledges funding from a Clean Energy Institute Graduate Research Fellowship. P.C. acknowledges the Swiss National Science Foundation for an Early Postdoc Mobility Fellowship (P2GEP2_148356). REFERENCES (1) Moon, G. D.; Ko, S.; Min, Y.; Zeng, J.; Xia, Y.; Jeong, U. Chemical Transformations of Nanostructured Materials. Nano Today 2011, 6, 186−203. (2) Fayette, M.; Robinson, R. D. Chemical Transformations of Nanomaterials for Energy Applications. J. Mater. Chem. A 2014, 2, 5965−5978. (3) Son, D. H.; Hughes, S. M.; Yin, Y.; Alivisatos, A. P. Cation Exchange Reactions in Ionic Nanocrystals. Science 2004, 306, 1009− 1012. (4) Donegá, C. d. M. Synthesis and Properties of Colloidal Heteronanocrystals. Chem. Soc. Rev. 2011, 40, 1512−1546. (5) Li, H.; Zanella, M.; Genovese, A.; Povia, M.; Falqui, A.; Giannini, C.; Manna, L. Sequential Cation Exchange in Nanocrystals: Preservation of Crystal Phase and Formation of Metastable Phases. Nano Lett. 2011, 11, 4964−4970. (6) Gupta, S.; Kershaw, S. V.; Rogach, A. L. 25th Anniversary Article: Ion Exchange in Colloidal Nanocrystals. Adv. Mater. 2013, 25, 6923− 6944. (7) Beberwyck, B. J.; Surendranath, Y.; Alivisatos, A. P. Cation Exchange: A Versatile Tool for Nanomaterials Synthesis. J. Phys. Chem. C 2013, 117, 19759−19770. (8) Rivest, J. B.; Jain, P. K. Cation Exchange on the Nanoscale: An Emerging Technique for New Material Synthesis, Device Fabrication, and Chemical Sensing. Chem. Soc. Rev. 2013, 42, 89−96. (9) Hodges, J. M.; Kletetschka, K.; Fenton, J. L.; Read, C. G.; Schaak, R. E. Sequential Anion and Cation Exchange Reactions for Complete Material Transformations of Nanoparticles with Morphological Retention. Angew. Chem., Int. Ed. 2015, 54, 8567−8567. (10) Nedelcu, G.; Protesescu, L.; Yakunin, S.; Bodnarchuk, M. I.; Grotevent, M. J.; Kovalenko, M. V. Fast Anion-Exchange in Highly Luminescent Nanocrystals of Cesium Lead Halide Perovskites (CsPbX3, X = Cl, Br, I). Nano Lett. 2015, 15, 5635−5640. (11) Bryan, J. D.; Gamelin, D. R. Doped Semiconductor Nanocrystals: Synthesis, Characterization, Physical Properties, and Applications. Prog. Inorg. Chem. 2005, 54, 47−126. (12) Norris, D. J.; Efros, A. L.; Erwin, S. C. Doped Nanocrystals. Science 2008, 319, 1776−1779. (13) Beaulac, R.; Ochsenbein, S. T.; Gamelin, D. R. Colloidal Transition-Metal-Doped Quantum Dots. In Nanocrystal Quantum Dots, 2nd ed.; Klimov, V. I., Ed.; CRC Press: Boca Raton, FL, 2010; pp 397− 453. (14) Mocatta, D.; Cohen, G.; Schattner, J.; Millo, O.; Rabani, E.; Banin, U. Heavily Doped Semiconductor Nanocrystal Quantum Dots. Science 2011, 332, 77−81. (15) Buonsanti, R.; Milliron, D. J. Chemistry of Doped Colloidal Nanocrystals. Chem. Mater. 2013, 25, 1305−1317. (16) Schimpf, A. M.; Knowles, K. E.; Carroll, G. M.; Gamelin, D. R. Electronic Doping and Redox-Potential Tuning in Colloidal Semiconductor Nanocrystals. Acc. Chem. Res. 2015, 48, 1929−1937. (17) Erickson, C. S.; Bradshaw, L. R.; McDowall, S.; Gilbertson, J. D.; Gamelin, D. R.; Patrick, D. L. Zero-Reabsorption Doped-Nanocrystal Solar Concentrators. ACS Nano 2014, 8, 3461−3467. H

DOI: 10.1021/acsnano.5b07389 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano Semiconductors: Predictions and Observations. Phys. Rev. B: Condens. Matter Mater. Phys. 1984, 30, 4021−4023. (40) Furdyna, J. K. Diluted Magnetic Semconductors. J. Appl. Phys. 1988, 64, R29−R64. (41) Luo, Y. R. Comprehensive Handbook of Chemical Bond Energies; CRC Press: Boca Raton, FL, 2007. (42) Bacher, G.; Maksimov, A. A.; Schömig, H.; Kulakovskii, V. D.; Welsch, M. K.; Forchel, A. Monitoring Statistical Magnetic Fluctuations on the Nanometer Scale. Phys. Rev. Lett. 2002, 89, 127201. (43) Barcz, A.; Karczewski, G.; Wojtowicz, T.; Kossut, J. Manganese Diffusion in MBE-Grown Cd(Mn)Te Structures. J. Cryst. Growth 1996, 159, 980−984. (44) Gösele, U. M. Fast Diffusion in Semiconductors. Annu. Rev. Mater. Sci. 1988, 18, 257−282. (45) Casavola, M.; van Huis, M. A.; Bals, S.; Lambert, K.; Hens, Z.; Vanmaekelbergh, D. Anisotropic Cation Exchange in PbSe/CdSe Core/ Shell Nanocrystals of Different Geometry. Chem. Mater. 2012, 24, 294− 302. (46) Groeneveld, E.; Witteman, L.; Lefferts, M.; Ke, X.; Bals, S.; Van Tendeloo, G.; de Mello Donega, C. Tailoring ZnSe-CdSe Colloidal Quantum Dots via Cation Exchange: From Core/Shell to Alloy Nanocrystals. ACS Nano 2013, 7, 7913−7930. (47) Sadtler, B.; Demchenko, D. O.; Zheng, H.; Hughes, S. M.; Merkle, M. G.; Dahmen, U.; Wang, L.-W.; Alivisatos, A. P. Selective Facet Reactivity during Cation Exchange in Cadmium Sulfide Nanorods. J. Am. Chem. Soc. 2009, 131, 5285−5293. (48) Chan, E. M.; Marcus, M. A.; Fakra, S.; ElNaggar, M.; Mathies, R. A.; Alivisatos, A. P. Millisecond Kinetics of Nanocrystal Cation Exchange Using Microfluidic X-ray Absorption Spectroscopy. J. Phys. Chem. A 2007, 111, 12210−12215. (49) Jamil, N. Y.; Shaw, D. The Diffusion of Mn in CdTe. Semicond. Sci. Technol. 1995, 10, 952−958. (50) Qu, L.; Peng, X. Control of Photoluminescence Properties of CdSe Nanocrystals in Growth. J. Am. Chem. Soc. 2002, 124, 2049−2055. (51) Carbone, L.; Nobile, C.; De Giorgi, M.; Della Sala, F.; Morello, G.; Pompa, P.; Hytch, M.; Snoeck, E.; Fiore, A.; Franchini, I. R.; Nadasan, M.; Silvestre, A. F.; Chiodo, L.; Kudera, S.; Cingolani, R.; Krahne, R.; Manna, L. Synthesis and Micrometer-Scale Assembly of Colloidal CdSe/CdS Nanorods Prepared by a Seeded Growth Approach. Nano Lett. 2007, 7, 2942−2950. (52) Yu, W. W.; Peng, X. Formation of High-Quality CdS and Other IIVI Semiconductor Nanocrystals in Noncoordinating Solvents: Tunable Reactivity of Monomers. Angew. Chem., Int. Ed. 2002, 41, 2368−2371. (53) Piepho, S. B.; Schatz, P. N. Group Theory in Spectroscopy with Applications to Magnetic Circular Dichroism; Wiley: New York, 1983. (54) Beaulac, R.; Archer, P. I.; van Rijssel, J.; Meijerink, A.; Gamelin, D. R. Exciton Storage by Mn2+ in Colloidal Mn2+-Doped CdSe Quantum Dots. Nano Lett. 2008, 8, 2949−2953.

I

DOI: 10.1021/acsnano.5b07389 ACS Nano XXXX, XXX, XXX−XXX